Model Context Protocol moderate risk

Kubernetes

Kubernetes MCP server provides a flexible interface to interact with Kubernetes clusters, supporting read and write operations, Helm, and OpenTelemetry.

Connections & Capabilities

Connects To

GitHubGitLabPostgreSQLMySQLMongoDBDockerKubernetesFilesystemAWSGCPAzure

Capabilities

readwriteexecadmin

Quickstart

Install

npx kubernetes-mcp-server@latest

Config

{
  "mcpServers": {
    "kubernetes-mcp-server": {
      "command": "npx",
      "args": ["-y", "kubernetes-mcp-server@latest"]
    }
  }
}

Exposed MCP Tools (7)

safe

List Pods

Lists pods in a specified namespace or all namespaces.

Read-only operation, no modification of resources.

safe

Get Pod

Retrieves details for a specific pod in a namespace.

Read-only operation, no modification of resources.

critical

Delete Pod

Deletes a pod from a specified namespace.

Destructive operation, can cause service disruption.

safe

Show Logs

Displays logs for a pod in a specified namespace.

Read-only operation, no modification of resources.

high

Exec

Executes a command inside a pod.

Can execute arbitrary commands, potentially leading to security breaches.

high

Install Helm Chart

Installs a Helm chart in a specified namespace.

Helm charts can contain malicious code or misconfigurations.

high

Uninstall Helm Release

Uninstalls a Helm release from a specified namespace.

Can remove critical applications or services.

Safety Assessment

The Kubernetes MCP server offers flexibility but requires careful configuration. Running in read-only mode significantly reduces risk. Exercise caution when performing write operations or executing commands within pods.

Supports read-only mode to prevent accidental modifications.
Can be configured with specific RBAC permissions in Kubernetes.
Provides logging and monitoring capabilities for auditing.
Allows configuration via TOML files for controlled settings.

Allows write operations, potentially leading to unintended changes.
Improper configuration can expose sensitive resources.
Executing commands inside pods can introduce security risks.
Helm chart installations can introduce vulnerabilities if not properly vetted.